FacebookGoogleSun

CSE 249A -- Computer Architecture Support for Big Data

 

Instructor: Dean Tullsen

tullsen at cs dot ucsd dot edu

office hours by appointment

Andreas Prodromou will be helping me out with the course in a variety of ways.

This is a paper reading course.  We'll read 2 or more papers a week.  Most will be presented by students.

I expect students to be enrolled for 2 units, s/u. If you desperately need either more units or to be graded, then come talk to me. If you are enrolled for 2 units, you are expected to present on your turn, and read the papers (before class!) and participate in discussions when it is not your turn.

Class meets TuTh 3:30-4:50 in HSS 1106B

I want to be open to directions the students want to take with this class, so any suggestions for papers you want to read, or specific topics you want to explore, please send them to me!

Schedule:

January 5 Introduction   Andreas
January 7 Benchmarks:
" Clearing the Clouds: A Study of Emerging Scale-out Workloads on Modern Hardware ", ASPLOS, 2012.		 Dean
January 12  Benchmarks:
" BigDataBench: a Big Data Benchmark Suite from Internet Services ", HPCA 2014
 Rajib
January 14 BlueDBM: An Appliance for Big Data Analytics Francis
January 19 TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems Colin
January 21 Fast in-memory transaction processing using RDMA and HTM Kunal
January 26 DjiNN and Tonic: DNN as a Service and Its Implications for Future Warehouse Scale Computers Jae Hyun
January 28 A Reconfigurable Fabric for Accelerating Large-Scale Datacenter Services Yashank
February 2 DaDianNao: A Machine-Learning Supercomputer Harsha
February 4 Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding Colin
February 9 Architecting to Achieve a Billion Requests Per Second Throughput on a Single Key-Value Store Server Platform Minxuan
February 11 GraphLab: A New Framework For Parallel Machine Learning Yixing
February 16 Scaling Distributed Machine Learning with the Parameter Server Honam
February 18 The sharing architecture: sub-core configurability for IaaS clouds. ASPLOS 2014 Kunal
February 23 In-Memory Big Data Management and Processing: A Survey Francis
February 25 Facade: A Compiler and Runtime for (Almost) Object-Bounded Big Data Applications Honam
March 1 TABLA: A Unified Template-based Framework for Accelerating Statistical Machine Learning Yashank and Harsha
March 3 Towards Energy Proportionality for Large-Scale Latency-Critical Workloads Jae Hyun
March 8 Towards Sustainable In-Situ Server Systems in the Big Data Era Minxuan and Yixing
March 10 Neuromorphic Accelerators: A Comparison Between Neuroscience and Machine-Learning Approaches Rajib

 


Other Possible Papers. Please skim the paper for relevance to the class. Feel free to introduce other papers or topics.

 

Learning both Weights and Connections for Efficient Neural Networks, NIPS

Fast Support for Unstructured Data Processing: the Unified Automata Processor, Yuanwei Fang (University of Chicago), Tung T. Hoang (University of Chicago), Michela Becchi (University of Missouri), Andrew A. Chien (University of Chicago), Micro 2015

HRL: Efficient and Flexible Reconfigurable Logic for Near-Data Processing, Mingyu Gao and Christos Kozyrakis (Stanford University), HPCA 2016

Towards Sustainable In-Situ Server Systems in the Big Data Era, Chao Li (Shanghai Jiao Tong University/University of Florida), Yang Hu (University of Florida), Longjun Liu (Xi'an Jiaotong University/University of Florida), Juncheng Gu (University of Florida), Mingcong Song (University of Florida), Xiaoyao Liang (Shanghai Jiao Tong University), Jingling Yuan (Wuhan University of Technology), and Tao Li (University of Florida), ISCA 2015

NumaGiC: a Garbage Collector for Big Data on Big NUMA Machines, Lokesh Gidra (LIP6-INRIA/UPMC), G Thomas (SAMOVAR-Telecom SudParis), Julien Sopena (LIP6-UPMC/INRIA), Marc Shapiro (LIP6-INRIA/UPMC), and Dang Nhan Nguyen (Chalmers university of technology), asplos 2015

Facade: A Compiler and Runtime for (Almost) Object-Bounded Big Data Applications, Khanh Nguyen, Kai Wang, Yingyi Bu, Lu Fang, Jianfei Hu, and Guoqing (Harry) Xu (University of California, Irvine), asplos 2015

Towards Energy Proportionality for Large-Scale Latency-Critical Workloads, David Lo (Stanford University), Liqun Cheng (Google), Rama Govindaraju (Google), Luiz Barroso (Google), Christos Kozyrakis (Stanford University), ISCA 2014

Improving In-Memory Database Index Performance with Intel Transactional Synchronization Extensions, Tomas Karnagel (TU Dresden), Roman Dementiev (Intel), Ravi Rajwar (Intel), Konrad Lai (Intel), Thomas Legler (SAP AG), Benjamin Schlegel (TU Dresden), Wolfgang Lehner (TU Dresden), HPCA 2014